From: Jens Axboe <axboe@kernel.dk>
To: Dave Chinner <david@fromorbit.com>
Cc: Christopher Lameter <cl@linux.com>,
Christoph Hellwig <hch@lst.de>,
Vitaly Kuznetsov <vkuznets@redhat.com>,
Ming Lei <tom.leiming@gmail.com>,
linux-block <linux-block@vger.kernel.org>,
linux-mm <linux-mm@kvack.org>,
Linux FS Devel <linux-fsdevel@vger.kernel.org>,
"open list:XFS FILESYSTEM" <linux-xfs@vger.kernel.org>,
Dave Chinner <dchinner@redhat.com>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Ming Lei <ming.lei@redhat.com>
Subject: Re: block: DMA alignment of IO buffer allocated from slab
Date: Tue, 25 Sep 2018 09:44:54 -0600 [thread overview]
Message-ID: <3d63a42f-837a-4bf6-665a-c3a8c8cb46e8@kernel.dk> (raw)
In-Reply-To: <20180925074910.GB31060@dastard>
On 9/25/18 1:49 AM, Dave Chinner wrote:
> On Mon, Sep 24, 2018 at 12:09:37PM -0600, Jens Axboe wrote:
>> On 9/24/18 12:00 PM, Christopher Lameter wrote:
>>> On Mon, 24 Sep 2018, Jens Axboe wrote:
>>>
>>>> The situation is making me a little uncomfortable, though. If we export
>>>> such a setting, we really should be honoring it...
>
> That's what I said up front, but you replied to this with:
>
> | I think this is all crazy talk. We've never done this, [...]
>
> Now I'm not sure what you are saying we should do....
>
>>> Various subsystems create custom slab arrays with their particular
>>> alignment requirement for these allocations.
>>
>> Oh yeah, I think the solution is basic enough for XFS, for instance.
>> They just have to error on the side of being cautious, by going full
>> sector alignment for memory...
>
> How does the filesystem find out about hardware alignment
> requirements? Isn't probing through the block device to find out
> about the request queue configurations considered a layering
> violation?
Right now it isn't a stacked property, so answering the question
isn't even possible beyond "what does the top device require".
> What if sector alignment is not sufficient? And how would this work
> if we start supporting sector sizes larger than page size? (which the
> XFS buffer cache supports just fine, even if nothing else in
> Linux does).
If sector alignment isn't sufficient, then we'd need to bounce 512b
formats... But I don't want to over-design something that isn't
relevant to real life setups. I'm not aware of anything that needs
memory aligned to that degree.
> But even ignoring sector size > page size, implementing this
> requires a bunch of new slab caches, especially for 64k page
> machines because XFS supports sector sizes up to 32k. And every
> other filesystem that uses sector sized buffers (e.g. HFS) would
> have to do the same thing. Seems somewhat wasteful to require
> everyone to implement their own aligned sector slab cache...
>
> Perhaps we should take the filesystem out of this completely - maybe
> the block layer could provide a generic "sector heap" and have all
> filesystems that use sector sized buffers allocate from it. e.g.
> something like
>
> mem = bdev_alloc_sector_buffer(bdev, sector_size)
>
> That way we don't have to rely on filesystems knowing anything about
> the alignment limitations of the devices or assumptions about DMA
> to work correctly...
I like that idea, would probably also need a mempool backing for
certain cases.
--
Jens Axboe
next prev parent reply other threads:[~2018-09-25 15:44 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-19 9:15 block: DMA alignment of IO buffer allocated from slab Ming Lei
2018-09-19 9:41 ` Vitaly Kuznetsov
2018-09-19 10:02 ` Ming Lei
2018-09-19 11:15 ` Vitaly Kuznetsov
2018-09-20 1:28 ` Ming Lei
2018-09-20 3:59 ` Yang Shi
2018-09-20 6:32 ` Christoph Hellwig
2018-09-20 6:31 ` Christoph Hellwig
2018-09-21 13:04 ` Vitaly Kuznetsov
2018-09-21 13:05 ` Christoph Hellwig
2018-09-21 15:00 ` Jens Axboe
2018-09-24 16:06 ` Christopher Lameter
2018-09-24 17:49 ` Jens Axboe
2018-09-24 18:00 ` Christopher Lameter
2018-09-24 18:09 ` Jens Axboe
2018-09-25 7:49 ` Dave Chinner
2018-09-25 15:44 ` Jens Axboe [this message]
2018-09-25 21:04 ` Matthew Wilcox
2018-09-23 22:42 ` Ming Lei
2018-09-24 9:46 ` Andrey Ryabinin
2018-09-24 14:19 ` Bart Van Assche
2018-09-24 14:43 ` Andrey Ryabinin
2018-09-24 15:08 ` Bart Van Assche
2018-09-24 15:52 ` Andrey Ryabinin
2018-09-24 15:58 ` Bart Van Assche
2018-09-24 16:07 ` Andrey Ryabinin
2018-09-24 16:19 ` Bart Van Assche
2018-09-24 16:47 ` Christopher Lameter
2018-09-24 18:57 ` Matthew Wilcox
2018-09-24 19:56 ` Bart Van Assche
2018-09-24 20:41 ` Matthew Wilcox
2018-09-24 20:54 ` Bart Van Assche
2018-09-24 21:09 ` Matthew Wilcox
2018-09-25 0:16 ` Ming Lei
2018-09-25 3:28 ` Matthew Wilcox
2018-09-25 4:10 ` Bart Van Assche
2018-09-25 4:44 ` Matthew Wilcox
2018-09-25 6:55 ` Ming Lei
2018-09-24 15:17 ` Christopher Lameter
2018-09-25 0:20 ` Ming Lei
2018-09-20 14:07 ` Bart Van Assche
2018-09-21 1:56 ` Dave Chinner
2018-09-21 7:08 ` Christoph Hellwig
2018-09-21 7:25 ` Ming Lei
2018-09-21 14:59 ` Jens Axboe
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3d63a42f-837a-4bf6-665a-c3a8c8cb46e8@kernel.dk \
--to=axboe@kernel.dk \
--cc=cl@linux.com \
--cc=david@fromorbit.com \
--cc=dchinner@redhat.com \
--cc=hch@lst.de \
--cc=linux-block@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-xfs@vger.kernel.org \
--cc=ming.lei@redhat.com \
--cc=tom.leiming@gmail.com \
--cc=vkuznets@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).